Diseases of the central nervous system (CNS) are a significant public health and economic problem, affecting nearly one in three Americans at some point in life, with a cost exceeding $500 billion per year. Pathologically, the axonal integrity in the brain white matter is compromised in most human neurological diseases or injuries, including multiple sclerosis, leukodystrophy, traumatic brain injury, as well as in normal aging and neurodegenerative diseases such as Parkinson's, Huntington's, amyotrophic lateral sclerosis, and Alzheimer's disease. Neurodegeneration occurs via a number of distinct processes, such as acute axonal injury, demyelination, inflammation, chronic axonal degeneration or axonal loss, and gliosis. Quantifying their degree, and distinguishing between these different underlying degenerative processes is particularly vital for adequate assessment of CNS disease progression and treatment, but currently not available. The main objective of this proposal is the development, validation and clinical translation of MesoMRI, a newly proposed framework that yields high specificity of diffusion MRI metrics to different degenerative processes by identifying how structural changes at the mesoscopic scale manifest themselves in the diffusion-weighted signal. The mesoscopic scale is the scale of cellular tissue architecture, intermediate between the molecular scale and the macroscopic MRI resolution. Our overall hypothesis is that our novel mesoscopic biomarkers of brain tissue integrity will be specific to different aspects of neurodegeneration in vivo, including demyelination and axonal loss. The development of the mesoscopic MRI framework involves qualitative understanding and quantitative modeling of the role of axonal geometry at the micrometer scale, and the response of water diffusion metrics to demyelination, axonal loss, injury, and other pathological changes. This involves bringing advanced analytical methods borrowed from modern transport theory and statistical physics into the context of diffusion MRI. In particular, we will first deveop and validate our white matter tract integrity metrics and the relation between them and the degree of demyelination and axonal loss in a single white matter fiber bundle. Subsequently, we will extend the mesoscopic modeling of a single fiber bundle onto the whole white matter, including regions of multiple fiber directions and crossings. This extension involves the development of MesoFT, a novel mesoscopic modeling and fiber tracking paradigm, able for the first time to produce self-consistent maps of both local mesoscopic parameters of fibers and their macroscopic connectivity maps. The potential of our technical developments for clinical translation will be evaluated by applying MesoFT to retrospectively acquired clinical cases of Alzheimer's disease and multiple sclerosis. Based on adequate interpretation of standard clinically feasible diffusion MRI metrics, our framework's translation into clinic will be straightforward, and will enable the quantitative assessment of disease progression and quality of treatment in neurological diseases.